Anna Puig | University of Barcelona (original) (raw)

Papers by Anna Puig

Research paper thumbnail of An approach to improve user experience with conversational recommenders through a 3D virtual environment

Proceedings of the XV International Conference on Human Computer Interaction - Interacción '14, 2014

Research paper thumbnail of A Generic Framework to Exploit Virtual Worlds as Normative and Dynamic Interactive Spaces

In this research we present a generic framework for behavior management in social virtual worlds.... more In this research we present a generic framework for behavior management in social virtual worlds. Due to the type of activities taking place in a social virtual world, it is important to rely on mechanisms ensuring that the virtual environment is prepared to be a dynamic space where participants are informed about activities evolution and where norms are used to organize participants' actions, to define actions' consequences and to prevent undesired participants behaviours. We define an interaction framework where model rendering and event capture are separated from decision mechanisms allowing to be used by an IA based module (e.g multiagent system) and exploited by different virtual world platforms. We present a prototype developed using Wonderland platform where we contribute with a new type of component named iObjectCell and a new scheme of multi-view model.

Research paper thumbnail of Iterative multi-class multi-scale stacked sequential learning: Definition and application to medical volume segmentation

Pattern Recognition Letters, 2014

ABSTRACT In this work we present the iterative multi-class multi-scale stacked sequential learnin... more ABSTRACT In this work we present the iterative multi-class multi-scale stacked sequential learning framework (IMMSSL), a novel learning scheme that is particularly suited for medical volume segmentation applications. This model exploits the inherent voxel contextual information of the structures of interest in order to improve its segmentation performance results. Without any feature set or learning algorithm prior assumption, the proposed scheme directly seeks to learn the contextual properties of a region from the predicted classifications of previous classifiers within an iterative scheme. Performance results regarding segmentation accuracy in three two-class and multi-class medical volume datasets show a significant improvement with respect to state of the art alternatives. Due to its easiness of implementation and its independence of feature space and learning algorithm, the presented machine learning framework could be taken into consideration as a first choice in complex volume segmentation scenarios.

Research paper thumbnail of Hybrid Model for Vascular Tree Structures

Eurographics, 2000

This paper proposes a new representation scheme of the cerebral blood vessels. This model provide... more This paper proposes a new representation scheme of the cerebral blood vessels. This model provides information on the semantics of the vascular structure: the topological relationships between vessels and the labeling of vascular accidents such as aneurysms and stenoses. In addition, the model keeps information of the inner surface geometry as well as of the vascular map volume properties, i.e. the tissue density, the blood ow velocity and the vessel wall elasticity. The model can be constructed automatically in a pre-process from a set of segmented MRA images. Its memory requirements are optimized on the basis of the sparseness of the vascular structure. It allows fast queries and e cient traversals and navigations. The visualizations of the vessel surface can be performed at di erent levels of detail. The direct rendering of the volume is fast because the model provides a natural way to skip over empty data. The paper analyzes the memory requirements of the model along with the costs of the most important operations on it.

Research paper thumbnail of Classification Algorithms for Biomedical Volume Datasets

Lecture Notes in Computer Science, 2006

Abstract. This paper analyzes how to introduce machine learning algorithms into the process of di... more Abstract. This paper analyzes how to introduce machine learning algorithms into the process of direct volume rendering. A conceptual framework for the optical property function elicitation process is proposed and particularized for the use of attribute-value classifiers. The process is evaluated in terms of accuracy and speed using four different off-theshelf classifiers (J48, Naıve Bayes, Simple Logistic and ECOC-Adaboost). The empirical results confirm the classification of biomedical datasets as a tough problem where an opportunity ...

Research paper thumbnail of Discrete Medial Axis Transform for Discrete Thin and Elongated Objects

Abstract: Skeletons are compact shape descriptions of discrete images. They have been extensively... more Abstract: Skeletons are compact shape descriptions of discrete images. They have been extensively studied because of their utility in various applications such as data compression, shape abstraction, navigation and features detection. In this article, a new Euclidean skeletal definition for 2D discrete objects based on the Distance Map (DMAT) is being proposed. As a novel feature, it is shown that this skeleton is a connected subset of the discretization of the continuous medial axis of the object....

Research paper thumbnail of Visualization and Data Analysis 2006 (Proceedings Volume)

We present MIBFV, a method to produce real-time, multiscale animations of flow datasets. MIBFV ex... more We present MIBFV, a method to produce real-time, multiscale animations of flow datasets. MIBFV extends the attractive features of the Image-Based Flow Visualization (IBFV) method, ie dense flow domain coverage with flow-aligned noise, real-time animation, implementation simplicity, and few (or no) user input requirements, to a multiscale dimension. We generate a multiscale of flow-aligned patterns using an algebraic multigrid method and use them to synthesize the noise textures required by IBFV. We demonstrate our approach with ...

Research paper thumbnail of An approach to improve user experience with conversational recommenders through a 3D virtual environment

Proceedings of the XV International Conference on Human Computer Interaction - Interacción '14, 2014

Research paper thumbnail of A Generic Framework to Exploit Virtual Worlds as Normative and Dynamic Interactive Spaces

In this research we present a generic framework for behavior management in social virtual worlds.... more In this research we present a generic framework for behavior management in social virtual worlds. Due to the type of activities taking place in a social virtual world, it is important to rely on mechanisms ensuring that the virtual environment is prepared to be a dynamic space where participants are informed about activities evolution and where norms are used to organize participants' actions, to define actions' consequences and to prevent undesired participants behaviours. We define an interaction framework where model rendering and event capture are separated from decision mechanisms allowing to be used by an IA based module (e.g multiagent system) and exploited by different virtual world platforms. We present a prototype developed using Wonderland platform where we contribute with a new type of component named iObjectCell and a new scheme of multi-view model.

Research paper thumbnail of Iterative multi-class multi-scale stacked sequential learning: Definition and application to medical volume segmentation

Pattern Recognition Letters, 2014

ABSTRACT In this work we present the iterative multi-class multi-scale stacked sequential learnin... more ABSTRACT In this work we present the iterative multi-class multi-scale stacked sequential learning framework (IMMSSL), a novel learning scheme that is particularly suited for medical volume segmentation applications. This model exploits the inherent voxel contextual information of the structures of interest in order to improve its segmentation performance results. Without any feature set or learning algorithm prior assumption, the proposed scheme directly seeks to learn the contextual properties of a region from the predicted classifications of previous classifiers within an iterative scheme. Performance results regarding segmentation accuracy in three two-class and multi-class medical volume datasets show a significant improvement with respect to state of the art alternatives. Due to its easiness of implementation and its independence of feature space and learning algorithm, the presented machine learning framework could be taken into consideration as a first choice in complex volume segmentation scenarios.

Research paper thumbnail of Hybrid Model for Vascular Tree Structures

Eurographics, 2000

This paper proposes a new representation scheme of the cerebral blood vessels. This model provide... more This paper proposes a new representation scheme of the cerebral blood vessels. This model provides information on the semantics of the vascular structure: the topological relationships between vessels and the labeling of vascular accidents such as aneurysms and stenoses. In addition, the model keeps information of the inner surface geometry as well as of the vascular map volume properties, i.e. the tissue density, the blood ow velocity and the vessel wall elasticity. The model can be constructed automatically in a pre-process from a set of segmented MRA images. Its memory requirements are optimized on the basis of the sparseness of the vascular structure. It allows fast queries and e cient traversals and navigations. The visualizations of the vessel surface can be performed at di erent levels of detail. The direct rendering of the volume is fast because the model provides a natural way to skip over empty data. The paper analyzes the memory requirements of the model along with the costs of the most important operations on it.

Research paper thumbnail of Classification Algorithms for Biomedical Volume Datasets

Lecture Notes in Computer Science, 2006

Abstract. This paper analyzes how to introduce machine learning algorithms into the process of di... more Abstract. This paper analyzes how to introduce machine learning algorithms into the process of direct volume rendering. A conceptual framework for the optical property function elicitation process is proposed and particularized for the use of attribute-value classifiers. The process is evaluated in terms of accuracy and speed using four different off-theshelf classifiers (J48, Naıve Bayes, Simple Logistic and ECOC-Adaboost). The empirical results confirm the classification of biomedical datasets as a tough problem where an opportunity ...

Research paper thumbnail of Discrete Medial Axis Transform for Discrete Thin and Elongated Objects

Abstract: Skeletons are compact shape descriptions of discrete images. They have been extensively... more Abstract: Skeletons are compact shape descriptions of discrete images. They have been extensively studied because of their utility in various applications such as data compression, shape abstraction, navigation and features detection. In this article, a new Euclidean skeletal definition for 2D discrete objects based on the Distance Map (DMAT) is being proposed. As a novel feature, it is shown that this skeleton is a connected subset of the discretization of the continuous medial axis of the object....

Research paper thumbnail of Visualization and Data Analysis 2006 (Proceedings Volume)

We present MIBFV, a method to produce real-time, multiscale animations of flow datasets. MIBFV ex... more We present MIBFV, a method to produce real-time, multiscale animations of flow datasets. MIBFV extends the attractive features of the Image-Based Flow Visualization (IBFV) method, ie dense flow domain coverage with flow-aligned noise, real-time animation, implementation simplicity, and few (or no) user input requirements, to a multiscale dimension. We generate a multiscale of flow-aligned patterns using an algebraic multigrid method and use them to synthesize the noise textures required by IBFV. We demonstrate our approach with ...